The rumour mill is in high gear today, but it’s a rumour coming from kopite7kimi, who’s hit the mark on rumours a good few times already. The information says that the new Nvidia Ampere RTX 3000 series graphics cards will be made on the Samsun 8nm process, and not the TSMC 7nm node that we thought it was. Of course, AMD is using 7nm, and that could give them an edge as the platform matures.
Nvidia RTX 3000 Ampere
What we do know, however, is that Samsungs 8nm 8LPP is developed from their 10nm 10LPP process, similar to that used on the Galaxy phones. However, the process was highly optimised and formed the basis for the 8nm tech, allowing for a transistor density of 71 million/mm2. This 8LPP tech is expected to be the power station for Nvidia GA102 silicon that will power their three flagship cards.
Despite having a less advanced process, Nvidia is well known for being able to produce more efficient cards, and great clock speeds as a result. The RTX 2080 Ti is 12nm, and that’s no slouch. Of course, how it all stacks up remains to be seen, as we’re yet to see AMD put a Big Navi card on the market.
100% Samsung 8nm(develop from 10nm) https://t.co/jYqkSYfku7
— kopite7kimi (@kopite7kimi) July 2, 2020
What We Think
Honestly, this rumour could go either way, as we did expect Nvidia to match AMD at 7nm using TSMC tech. However, the 8nm rumour does line up with the other rumour of higher power usage than expected. Nvidia’s cards would use more power at 8nm than 7nm to get the performance figures expected.
